CG-AL-X278

hard Codeunits & Business Logic content b70b0581637c…

Description

Per-model results

Models that have attempted this task
ModelAttempt 1Attempt 2Avg scoreTokens / attemptRuns
Claude Opus 5✗✓81.3 / 10012k / 17k3
Claude Opus 5.5✗✓81.3 / 10012k / 4.7k3
GPT-5.6 Sol✗✓81.3 / 1007.6k / 4.8k3
GPT-6 Sol✗✓81.3 / 1007.1k / 2.9k3
Claude Sonnet 5✗✓75.0 / 10013k / 17k3
GPT-6.1 Sol✗✓75.0 / 1006.7k / 2.8k3
GPT-5.6 Terra✗✓68.8 / 1007.5k / 3.4k3
GPT-6 Luna✗✓68.8 / 1006.8k / 4.4k3
Gemini 3.8 Flash (OpenRouter)✗✓68.8 / 1009.1k / 35k3
Claude Haiku 4 5✗✗62.5 / 10010k / 1.9k3
GPT-5.6 Luna✗✓58.3 / 1007.3k / 3.2k3
MiniMax M3✗✓58.3 / 1007.4k / 17k3
GLM 5.3 (OpenRouter)✗✗31.3 / 1007.3k / 59k2